https://github.com/krizhanovsky/NatSys-Lab
编译:g++ -Wall -std=c++11 -O2 -D DCACHE1_LINESIZE=`getconf LEVEL1_DCACHE_LINESIZE` lockfree_rb_q.cc -lpthread -o test
Conclusion
Nowadays, high-performance computing typically is achieved in two ways: horizontal scaling (scale-out) by adding new computational nodes and vertical scaling (scale-up) by adding extra computational resources (like CPUs or memory) to a single node. Unfortunately, linear scaling is possible only in theory. In practice, if you double your computational re